Letter Frequency Analyzer - Character Frequency Counter

A letter frequency analyzer counts how many times each character or letter appears in a text and converts the counts into percentages and visual charts. It is used for writing analysis, linguistics, education, and cryptanalysis to understand distribution patterns like which letters dominate English (E, T, A, O, I, N).

Percentage is calculated as count / total counted tokens * 100. For example, if e appears 12 times out of 100 filtered letters, its frequency is 12%. When you use bigrams or enable filters, the total is the number of tokens after filtering, not the raw input length. Bar width is relative to the most frequent token.

In case-insensitive mode (default), A and a are merged using toLocaleLowerCase() before counting, giving you the classic A-Z distribution. In case-sensitive mode, uppercase and lowercase are counted separately, which is useful for analyzing code, passwords, or stylized text where case carries meaning.

Shannon Entropy in bits measures randomness: -Σ p·log₂(p). Low entropy (~3-4 bits) means predictable, repetitive text. High entropy (~4.5+ bits for letters) means more uniform distribution. Index of Coincidence measures the probability that two random letters are identical. Natural English is about 0.0667, random text is about 0.0385. Cryptanalysts use IoC to detect if text is encrypted or to guess key length.

Yes. Counting uses Unicode-aware Array.from() and regex properties \p{L} and \p{N}, so accented letters, Cyrillic, Arabic, Chinese characters, and emojis are counted as single characters. Turn off Unicode: on to restrict analysis to ASCII only (code points 0-127). Display names like ␣ SPACE and ↵ LF make whitespace visible.

An n-gram is a contiguous sequence of N tokens. 1-gram = single characters, bigram = 2 characters like th, he, trigram = 3 characters like the, ing. Analyzing bigrams and trigrams reveals writing style, common phonetic patterns, and is essential for breaking ciphers and building language models. The tool builds n-grams after filtering, so spaces are excluded unless you include them.

In English, E is most frequent (~12.7%), followed by T, A, O. In a simple substitution cipher, letter frequencies are preserved but remapped. Paste the ciphertext, set Letters only and Case insensitive, enable English reference, and compare your top bars to the amber expected lines. If your top cipher letter is K, it likely maps to E. Repeat with bigrams (th, he are most common) to confirm.

No. The entire analysis runs 100% in your browser using JavaScript. No text is sent to a server, stored in localStorage, or logged. You can disconnect from the internet and it will still work. Exports (CSV/JSON) are generated locally via Blob URLs.